Re: About boot logo

Pierre Phaneuf (pphaneuf@sx.nec.com)
Thu, 14 May 1998 11:42:03 -0400


Brian Weiss wrote:

> Instead of arguing about whether a boot logo is good or not, we should
> concentrate on how it should be done for many reasons.
>
> Most likely, it's going to happen in some form, and done right, it could
> be a very good thing. This fighting over it is getting us nowhere, and
> it's going to happen or not, regardless of all this.

I agree with you 100%.

To those that are against the logo: it *is* going to happen and if you
don't like it, just don't apply the patch, or don't set CONFIG_BOOTLOGO
(or whatever). I'm *pretty sure* Linus won't ever let a boot logo patch
in if it cannot be disabled completely, and if there's a patch which
puts one that cannot be disabled, well DON'T APPLY IT THEN!

> If the boot logo (or splash screen even I guess) is done right and
> intelligently...it should be a good thing for all types of users.
> Hopefully no one will lose any information they obtain from the boot
> msgs, but people wont be "scared" by them either.
>
> I'm personally in favor of a boot logo that still shows the boot msgs,
> but apparently most people want an option to completely turn them off or
> at least tone them down a bit. If this is the case with the boot
> logo/splash screen, It should definately be an option that is easily
> changed and would probably be best done from LILO. Of course, that does
> present the problem of people using another boot loader to boot Linux. I
> don't know what could be done about that.

There seems to be a few different things being discussed: a boot logo
(graphical or not), some layout/coloring of the boot messages and
changes to the boot messages themselves.

I can see those things be separate options, having a full screen splash
graphic, or a boot sequence with some panels (including (or not) a boot
logo), with "user-friendly" boot messages (or the regular ones), colored
or not.

Linux is about that, having the CHOICE.

> Also, there could be major problems with anything involving real
> graphics, such that it would require something similar to svgalib. The
> first reason for this is that Linux (graphics especially) doesn't
> usually just *work* right off the bat like win95 does somewhat. It
> usually requires at least some configuring and fixing up. Due to this
> it's safe to assume that this boot logo or splash screen would not just
> work for a lot of people. And since it would run before anything else it
> would increase the chances of things going wrong. The last thing Linux
> needs is problems with the booting itself :). Probably the safest way
> (at least for now) to do something like this would just to have a small
> logo that is displayed like a simple ansi logo.

This is technicality. I don't like much getting graphic support in
there, and if there's anything at all, I remember switching to mode 13
and blitting a raw image right into the video memory. The fancy version
of this includes 768 bytes at the beginning of the data to define the
palette instead of using the default palette.

If you don't have a VGA adapter, too bad. I guess that someone using a
monochrome monitor or some other old stuff for display probably isn't
all that picky about some splash screen...

Keeping in text mode, you still can do nice things. I remember some OS
called TSX-32 for PCs. It started in a way similar to loading Linux
using loadlin, and had this nice blue background, with a single window
that had the boot messages scrolling by, the name, copyright and version
of the OS at the top. You could have a simple screen dump (à la
/dev/vcsa*), add some window information and compile it in the virtual
console driver. You could add a function to set/unset "boot mode", maybe
having an ioctl interface on the /dev/vcs* devices, which would enable
blitting the "boot screen" to the video memory and setting up a window
of the specified size. Console message don't do anything fancy (they
don't assume much about the output device, it could be a serial printer
for all it cares), so they'd just print happily in the smaller window.
Unsetting this mode would clear the screen and use the full screen
again.

It would be nice to have this screen enabled/disabled by a user mode
program, so that you could let the kernel leave the screen in this state
after starting init, and have the very last script run this command,
right before starting up the getty's.

-- 
Pierre Phaneuf
Web: http://newsfeed.sx.nec.com/~ynecpip/
finger: ynecpip@newsfeed.sx.nec.com

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu